The presence of a chromosomal translocation in the leukemic cells at diagnosis of acute lymphoblastic leukemia (ALL) in children is associated with a high risk for treatment failure. We have reexamined the relationship between translocations and prognosis in 146 children with ALL who received risk-based therapy such that high-risk patients were treated with intensive drug schedules. In univariate analysis, multiple factors were associated with a relatively poor event-free survival (EFS) including age less than 2 years or greater than 10 years (combined group), WBC count greater than 10 x 10(9)/L, French-American-British (FAB) morphologic classification L2, absence of common ALL antigen (CALLA, CD10) expression, absence of hyperdiploidy with a chromosome number of 50 to 60, and presence of the specific translocations t(4; 11)(q21;q23) or t(9;22)(q34;q11) (combined group). However, there was no disadvantage with respect to EFS in patients with translocations compared with those who lacked translocations (73% at 4 years in both groups). Furthermore, when patients with specific cytogenetic abnormalities for which the prognostic significance has been well established (hyperdiploid 50 to 60, t(4;11), and t(9;22] were removed from the analysis, the remaining group with other translocations had a better EFS than the remaining group lacking translocations, although this was not statistically significant (81% v 65% at 4 years, P = .24). In a multivariate analysis, a model including WBC count and FAB classification was the strongest predictor of EFS. The presence or absence of translocations was not an independent predictor of EFS and did not contribute to the ability of any model to predict EFS. In conclusion, when effective intensive therapy is used to treat childhood ALL with high-risk clinical features, categorization of patients on the basis of chromosomal translocations without attention to the specific abnormality is not useful as a prognostic factor.